Key takeaways
- Weight every deal by the historical chance that a deal at its stage closes within the forecast window, then add them up.
- Pending deals are the most reliable stage; nationally, pending activity parallels closings over the following two months[1].
- Discount for fall-through: 47% of US sellers had at least one offer fall through in Zillow's 2026 survey[2].
- Publish a base, low and high case, and update the forecast weekly from the CRM.
Why forecast at all
Agents and brokers make money decisions months before the commission arrives. Whether to hire a coordinator, commit to a marketing contract, or set aside more for taxes all depends on what will close in the next quarter. Most people answer that question by looking at their pending list and hoping. A simple weighted forecast does much better, and it takes less than an hour to set up once your CRM stages are clean.
The goal is not precision. Real estate is lumpy, and one large deal can move a month. The goal is a range you can plan against, built from your own history rather than from optimism, and updated often enough to catch problems early.
Define the stages
- Active lead: a person in conversation with you about buying or selling within a year.
- Appointment held: you have met and discussed their plans.
- Signed client: a written buyer agreement or listing agreement is in place.
- Active listing or active search: the home is on the market, or the buyer is touring.
- Under contract: an accepted offer, with a scheduled closing date.
- Cleared to close: contingencies removed and the loan approved, if financed.
Each deal must sit in exactly one stage, with the date it entered that stage. Without entry dates you cannot measure how long deals take to move, which is half of the forecast.
Agree the stage definitions with every agent before the first forecast. If one agent marks a buyer as an active search after a single phone call and another waits for a signed agreement and a first tour, the probabilities will be meaningless.
The weighted pipeline
Expected closings in a window = the sum, across all stages, of deals at that stage × the historical probability that a deal at that stage closes within the window. Expected GCI = the same sum using each deal's expected commission.
The probabilities come from your own records. A deal under contract might have a high chance of closing in the next 60 days, while a signed listing might have a moderate chance and an active lead a small one. Multiply, add, and you have an expected number that already accounts for the deals that will not make it.
Find your stage probabilities
Pull 12 to 18 months of history
Export every deal with the dates it entered each stage and, if it closed, the closing date.
Pick the windows
Most agents use 30, 60 and 90 days; brokers often add 180 days for planning.
Count the outcomes
For each stage, find the share of deals that closed within each window after entering that stage.
Check the sample
If a stage has fewer than about 20 deals, combine it with the next stage or use a team or brokerage figure instead.
Refresh quarterly
Recalculate the probabilities every quarter so they reflect the current market.
Example: a 90-day forecast
| Stage | Deals now | Chance of closing within 90 days | Expected closings | Expected GCI |
|---|---|---|---|---|
| Cleared to close | 3 | 97% | 2.9 | $29,100 |
| Under contract | 9 | 85% | 7.7 | $76,500 |
| Active listing or search | 14 | 35% | 4.9 | $49,000 |
| Signed client, not yet active | 8 | 15% | 1.2 | $12,000 |
| Appointment held | 20 | 4% | 0.8 | $8,000 |
| Total | 54 | Weighted | 17.5 | $174,600 |
Example reading: the team should plan around 17 or 18 closings in the next 90 days, not the 26 deals that are under contract or actively listed. Most of the expected income comes from the first two stages; the earlier stages add a meaningful amount but carry most of the uncertainty.
Timing: when, not just whether
A deal can be likely to close and still miss the window. For deals under contract, use the closing date in the contract, then adjust for your history of delays. For earlier stages, use the typical time from that stage to closing, measured from your records. Nationally, NAR notes that pending sales activity parallels closed existing-home sales in the following two months[1], which is a reasonable starting assumption for the under-contract stage.
Earlier stages take much longer. Zillow found the median seller thought seriously about selling for three to four months before listing[2]. A seller at the appointment stage today may not list until next season, so do not drop them into this quarter's forecast just because the meeting went well.
Adjust for fall-through
Stage probabilities based on your history already include fall-through, but only at the rate you saw in the past. When conditions change, adjust. Zillow's 2026 survey found 47% of US sellers had at least one offer fall through during their sale, and financing problems were the most common reason[2]. If rates have risen since your deals went under contract, lower the probability for financed deals that have not cleared underwriting, and tell the agents why.
Adjust for market and season
Pipeline forecasts lean on history, so they miss turns in the market unless you look outward too. Zillow's preliminary data showed newly pending listings down 16.8% year over year in October 2026 and existing-home sales down 8.7%[3]. When new contracts fall that sharply, the pipeline of future closings thins out a month or two later, even if your current pending list looks healthy.
Seasonality matters as much. Realtor.com's research on the best week to list shows how much spring timing can change results: listings that week have historically drawn more views, sold faster and fetched higher prices than the average week[4]. Weight your stage-to-close timing by season, using the same quarter in previous years.
Publish three cases
- Base case: the weighted forecast as calculated.
- Low case: lower the probabilities for the two earliest stages by a third and push a share of under-contract deals past their closing dates.
- High case: add the most likely early-stage deals at higher probabilities, if the market is improving.
- Plan fixed costs and hiring against the low case, and marketing and growth spending against the base case.
Example: applied to the fictional team above, the low case might come to about 14 closings and the high case about 20, around a base of 17 or 18. The team leader commits to fixed costs that the low case can cover, such as keeping the coordinator on the current hours, and holds back a planned advertising increase until the next update shows whether the base case is holding. Sharing all three numbers with agents also sets honest expectations: the forecast is a range, and the weekly pipeline meeting is where it narrows.
Rolling it up for a brokerage
Brokers can add up agent forecasts, but agent-level probabilities are often too optimistic and samples too small. A better approach uses brokerage-wide stage probabilities applied to every agent's deals, then compares the result with what agents say they expect. Large gaps between the two are worth a conversation: either the agent knows something the history does not, or the pipeline is being read hopefully.
Convert the forecast into company dollar as well as GCI by applying each agent's split, because capped agents add closings with little brokerage income. That company dollar forecast, set against fixed costs, is what tells a broker whether next quarter needs a cash buffer.
Keep it current
A forecast is only as good as the last update. Review it weekly in the pipeline meeting: move deals between stages, add new ones, remove dead ones and note any deal whose closing date moved. Once a month, compare last month's forecast with what actually closed. If the forecast keeps running high, your probabilities are optimistic or deals are sitting in stages too long.
The forecast connects naturally to the market data you read each month. Pair it with your MLS market report review and track buyer progress with showing-to-offer ratios. For the long-term value behind each client in the pipeline, see lifetime value of a real estate client.
Common forecasting mistakes
- Counting every pending deal as a closing.
- Leaving dead deals in early stages for months, which inflates the expected total.
- Using one probability for all buyers, when cash and financed deals behave differently.
- Forecasting GCI without the brokerage split, or company dollar without caps.
- Ignoring the season, so a spring-built set of probabilities is applied unchanged in December.
- Never comparing forecast with actual results, so nobody learns whether the method works.
Frequently asked questions
How do you forecast real estate closings?
Multiply the deals at each pipeline stage by the historical probability that a deal at that stage closes within your window, then add the results. Adjust for timing, fall-through and season.
What share of pending deals actually close?
Use your own contract-to-close history. Nationally, 47% of sellers had at least one offer fall through in Zillow's 2026 survey[2], so do not assume every pending deal closes.
How far ahead can a pipeline forecast see?
Pending deals give a reliable view of the next 30 to 60 days; NAR notes pending activity parallels closings over the following two months[1]. Beyond 90 days, the forecast depends on earlier stages and is much less certain.
How often should the forecast be updated?
Weekly for deal movements, monthly for comparing forecast with actual results, and quarterly for recalculating stage probabilities.
References
National Association of REALTORS® (2026). Pending Home Sales Decreased 7.3% in September (opens in a new tab) ↑
Zillow Research (2026). Sellers: Results from the Zillow Consumer Housing Trends Report 2026 (opens in a new tab) ↑
Zillow Research (2026). Newly Pending Listings Fell 17% In October (Zillow October 2026 Market Report) (opens in a new tab) ↑
Realtor.com Economic Research (2026). The Best Time To List a Home: The Week of April 12-18 (2026 analysis) (opens in a new tab) ↑